The Value Of Brush Cytology And Combined With DNA Image Cytometry In Diagnosing Indeterminate Biliary Strictures

BackgroundBiliary strictures are common in clinical practice.Malignant strictures are mostly caused by cholangiocarcinoma,pancreatic cancer,ampullary carcinoma and gallbladder cancer.These cancers are highly aggressive with a high mortality rate and short survival time.Early detection is of crucial importance for clinical management.Differentiating benign from malignant biliary strictures(MBS)is quite challenging,although imaging modalities such as multi-detector computed tomography(MDCT),magnetic resonance cholangiopancreatography(MRCP)and endoscopic ultrasonography(EUS)have been developed during the past decades.Endoscopic transpapillary brush cytology during endoscopic retrograde cholangiopancreatography(ERCP)has recently been used to obtain histological evidence of MBS,because it is technically easy to be performed,requiring a short procedure time together with a low complication rate.Brush cytology achieves a high specificity but a low sensitivity.The reported sensitivities and specificities of brush cytology for cancer detection range from 18%to 80%and 97%to 100%,respectively.Endoscopic transpapillary forceps biopsy is another method to obtain histological evidence.Although compared with brush cytology,the technique is technically more difficult and time consuming,it can theoretically provide larger tissue samples to histopathological examination.However,the sensitivity of forceps biopsy is also inadequate in clinical practice,ranging from 30%to 88%.Many studies on brush cytology and forceps biopsies are based on only small sample sizes.Studies with large sample sizes are strongely needed to clarify the diagnostic yields of brush cytology.In some difficult cases,alternative sampling techniques are required to provide a definitive diagnosis that will aid to choose a treatment strategy.In recent years,the use of endoscopic ultrasound-guided fine needle aspiration(EUS-FNA/B)for obtaining histological evidence of pancreatic cancer has been increasing because of its high sensitivity.Several reports have evaluated the efficacy of the procedure for biliary lesions.Although the reported SE for EUS-FNA is superior to that of brush cytology,FNA/B is associated with peritoneal seeding of perihilar masses.EUS-FNA/B has several limitations as a routine clinical procedure for all biliary strictures.The practicing physician is facing with the dilemma that how to manage patients without histopathologic diagnosis who have high probability of cancer based on clinical presentation.Currently,there is no guidance strategy for such patients.One approach is to repeat the ERCP and brush cytology,but there is few researches to support this strategy.Recently,several reports have revealed that fluorescence in situ hybridization(FISH)can improve the accuracy of brush cytology.But it has not widely used in clinical area.Some authors have attempted to increase sensitivity and specificity by the use of grasping instruments.A computer-based digital analysis,DNA image cytometry(DNA-ICM),assesses quantitative chromosomal changes,referred to as DNA-ploidy,in the nucley of atypical cells in reference to normal DNA values determined with unequivocally benign nuclei in the same smear.This characteristic is considered a more accurate predictor of disease progression,risk of relapse and survival than clinical stage,histological grade orspecific molecular markers.The changes in DNA content and nuclear ploidy intumor cells can be detected through quantitative DNA image cytometry(DNA-ICM),which is thought to be a relatively simple and objective analytical technique becauseit can be performed during ERCP without any additional procedures or instruments.Our study aims to estimate the value of brush cytology and explore the accuracy of repeated brush cytology in patients with MBS.Moreover,we will estimate the diagnostic value of RC,DNA-ICM and the combination of these 2 methods for differentiating between benign and malignant biliary strictures.Part 1.The utility of repeat brush cytology for malignant biliary stricturesObjective1.To study the value of brush cytology for MBS.2.To estimate the utility of repeat brush cytology for MBS.3.To analyze the differences of the value among brush cytology?TPB and EUS-FNA for MBS.4.The association of diagnostic yield for MBS.Methods1.We collected patients who received Endoscopic transpapillary brush cytology from Changhai hospital(Shanghai)during 2013 and 2016.There are 476 cases including in our study.Furthermore,we evaluated various parameters to clarify the predictive factors for a positive diagnosis of MBS:sex/age/reason for strictures/CA199/TBIL/AST/ALT/?-GT/ALP.2.To perform ERCP and brush cytology.Some others perfom EUS-FNA or bile duct biopsy.The samples routinely send to cytological examination and pathologic diagnosis.3.Make sure the clariy of the diagnosis by elephone follow-up.Statistical analysis was performed using SPSS 19.0.Continuous variables were showed by the mean standard,and qualitative data are tested by chi-square test while quantitative data satisfying normal distribution are tested by t-test.Risk factors of malignant biliary stricture were analyzed by ROC regression.It is considered significant difference on P<0.05.Results1.The SE,SP,accuracy,NPV and PPV of brush cytology and repeat brush cytology were 30.29%vs 41.18%,100%vs 100%,55.04%vs 74.36%,51.6%vs 68.75%and 100%vs 100%,respectively.Repeated brush cytology has higher sensitivity,NPV and accuracy than single way,and the differences of NPV and accuracy is statistically significant.2.EUS-FNA and bile dust biopsy:the SE,SP,NPV PPV and accuracy,of EUS-FNA and bile duct biopsy were 44.44%vs 52.38%?100%vs 100%?100%vs 100%?28.57%vs 56.25%?54.55%74.07%.The diagnostic accuracy of EUS-FNA for pancreatic cancer and cholangiocarcinoma were 33.33%and 44.44%respectively.And the diagnostic accuracy of bile dust biopsy for pancreatic cancer and cholangiocarcinoma were 20%and 66.67%respectively.The diagnostic accuracy of EUS-FNA in pancreatic cancer was better that of bile dust biopsy.And the diagnostic accuracy of bile dust biopsy in pancreatic cancer was better that of EUS-FNA.However,the difference is not statistically significant.The diagnostic accuracy of brush cytology and repeat brush cytology for pancreatic cancer and cholangiocarcinoma were 25.6%vs 37.5%%and 33.3%vs 44.44%respectively.The accuracy of the brush cytology for pancreatic cancer and cholangiocarcinoma has significant difference.3.We compared the malignant and benign MBS patients base data who received repeat brush cytology.One-Way ANOVA shows that TBIL/length of stricture/ALT/CEA/CA199 were significant difference while the length of stricture and TBIL were the predictors of malignant by multivariate analysis.ROC regression analysis shows that length>22.5mm and TBIL>60.65mmol/L were related to the malignant biliary stricture.ConclusionIn conclusion,repeat brush cytology can improve the NPV and accuracy of MBS which would be wildly used in patients with stricture length>2.25cm and TBIL>60.65mmol/L who shows a negative result for the first time.Part2.The value of DNA image cytometry combined with brush routine cytology in diagnosing indeterminate biliarystricturesObjectiveThe study aimed to estimate the value of routine cytology(RC),DNA-ICM,and their combination in diagnosing indeterminate biliary strictures.Methods1.Materials:Patients with an indeterminate biliary stricture based on imaging examinations from January 2016 to January 2018 were retrospectively identified from theDepartment of Gastroenterology at Changhai Hospital at the Second/Navy Military Medical University in Shanghai,China.Moreover,the demographic information of thepatients(age and sex),imaging information(location of the stricture),follow-up imaging information(whether the stricture showed significant progress),cytopathologic examination results,and DNA-ICM and surgical pathology findings were obtained from our electronic medical system.2.To perform ERCP and brush cytology.The samples were divided into two parts:one for cytological examination and one for DNA-ICM examination.3.Brush cytology and DNA-ICM:We used cytological smears and liquid-based cytological methods to make a cytological diagnosis.After brushing the stricture,the brush was shaken for 10s or more in a vial containing normal saline.The vial was sent to Department of Pathology to perform a centrifugation.After centrifugation,the supernatant was abandoned and the sediment was directly smeared onto one or more glass slides.The conventional smears was immediately fixed using an alcohol-based fixative and stained according to the Papanicolaou(PAP)method.Then a cytologist comes to confirm the diagnosis through microscope.We used the Feulgen method to make a DNA-ICM diagnosiswith.The analysis of DNA contents and nuclear ploidy was performed by DNA-ICM using an automated cell screening and analysis system(Heer Medical Science and Technology Developing Co.,Ltd.Wuhan,China).4.Statistical analysisSPSS 20.0(IBM,Armonk,NY)was used for statistical analysis.All categorical variables are expressed as frequencies and proportions.The SE,specificity(SP),positive predictive value(PPV),negative predictive value(NPV),and accuracy of RC,DNA-ICM and the combination of these 2 methods were assessed along with 95%confidence intervals(CIs).The Chi-square test and area under the receiver operating characteristic(ROC)curve(AUC)analysis were used to compare the ability to detect malignancy with these 2 methods.P<0.05 was considered statistically significant.Results1.Cytology:Applying the specimen data to the diagnostic data,the SE,SP,accuracy,NPV,PPV and AUC of RC were 6.28%,100%,42.27%,39.94%,100%and 0.531,respectively,for all strictures.For proximal strictures,the SE,SP,accuracy,NPV,PPV and AUC of RC were 12.82%,100%,39.29%,33.33%,100%and 0.564,respectively,and 4.89%,100%,42.81%,41.08%,100%and 0.524,respectively,fordistal strictures.The diagnostic accuracy of RC for distal strictures was not better than that of RC for proximal strictures(P=0.623).2.DNA-ICM:the SE,SP,accuracy,NPV,PPV and AUC of DNA-ICM were 44.84%,92.8%,63.3%,51.2%,90.9%and 0.688,respectively.Of the 56 proximal stricture samples,The SE,SP,accuracy,NPV,PPV and AUC of DNA-ICM were 41.03%,76.47%,51.8%,36.11%,80%and 0.587,respectively.Of the 306 distal strictures,The SE,SP,accuracy,NPV,PPV and AUC of DNA-ICM were 45.65%,95.08%,65.36%,53.7%,93.33%and 0.703,respectively.Furthermore,the diagnostic accuracy of DNA-ICM for distal strictures was significantly higher than that of DNA-ICM for proximal strictures.3.Cytology and DNA-ICM:The SE,SP,accuracy,NPV,PPV and AUC of DNA-ICM combined with RC for proximal strictures were 46.15%,76.47%,55.36%,38.24%,81.82%and 0.6,respectively.Of the 306 cases of distal stricture,92 were positive for malignancy,and 204 were negative.The SE,SP,accuracy,NPV,PPV and AUC of DNA-ICM combined with RC in distal strictures were 46.74%,95.08%,66.01%,54.21%,93.48%and 0.738,respectively.4.Comparison of cytology and DNA-ICM:The diagnostic accuracy of DNA-ICM was significantly better than that of RC.However,in proximal strictures,DNA-ICM showed no superiority to RC.The diagnostic accuracy of DNA-ICM in distal strictures was better than that of RC.The combination of RC and DNA-ICM also had significantly better diagnostic accuracy than RC(64.36%vs 42.3%,P<0.001)in all strictures.Conclusion1.This study demonstrated that DNA-ICM was an objective and effective additional tool to improve the diagnostic accuracy of RC,and specimens for DNA-ICM can be easily obtained during ERCP.2.DNA-ICM was more accurate for distal than for proximal strictures.The combination of DNA-ICM and RC can also improve the diagnostic accuracy,especially for proximal strictures.